Research On CRM Based On Data Mining

Recent years, the competition between enterprises increasingly focuses on for the end-customers, customers become critical success factors and profit sources. So, Customer Relationship Management (CRM) that can provide an overall management of customer relationship and provide some related analysis become enterprises'focus. It is a key point in the information construction course.At first, introduce the fundamental theories, and then explain the meanings and characteristic of CRM in detail. After the analysis of relationship between CRM and Enterprise Resource Planning (ERP), we proposed an integrated solution of CRM and ERP to achieve a unified management of inside and outside enterprise resources to serve enterprise and customers better.CRM based on Data Mining is the kernel part. After an analysis of Data Mining technology and its application in CRM, we discussed the basic implement flow of Data Mining. The classification of customer is the basic and crucial link of implementation of CRM. Generally, to make a large customer base into small groups by clustering analysis techniques. The customers within same sub-group have similar characters. However, they have smaller similarities in different sub-groups. It guides enterprises to take differences of strategy to different sub-groups of customers on basis of this model.K-means algorithm, which based on division, is a classic cluster algorithm. Aiming at its selection mechanism of initial clustering centers, proposed an improved K-means algorithm. After tests, it is proved that the algorithm can get stable clustering results. That's to say it has overcome sensitive issue to the initial clustering centers. At last we pay attention to discuss the course of building a customer classification model based on customer value by K-means cluster algorithm. Through this process, customers can be grouped according to their contribution to enterprise. It can play an important guiding significance for maintaining high-value customers, resetting value or abandoning negative value customers.